Re: [Help-glpk] Table Simplex: inverse of the matrix of the basic variables

[Andrew Makhorin]

> I wonder how can I get (using API functions) the inverse of the basic
> matrix variables (B^-1) after obtaining the optimum of a linear
> problem.

The api routine glp_ftran computes y = inv(B) * x for a given vector x,
so if you specify x = ej, where ej is j-th unity vector (i.e. j-th
column of unity matrix), you can compute inv(B) * ej, which is j-th
column of inv(B). Thus, using glp_ftran you can compute entire inv(B) by
columns. Similarly, using glp_btran you can compute inv(B) by rows. For
details please see chapter "Advanced API Routines" in the GLPK Reference
Manual.